lable at ScienceDirect

Marine and Petroleum Geology 78 (2016) 222e235

Contents lists avai

Marine and Petroleum Geology

journal homepage: www.elsevier .com/locate/marpetgeo

Research paper

Permeability and frictional properties of halite-clay-quartz faults inmarine-sediment: The role of compaction and shear

Bryan M. Kaproth a, b, Marek Kacewicz c, Sankar Muhuri c, Chris Marone a, *

a Department of Geosciences, and Energy Institute Center for Geomechanics, Geofluids, and Geohazards, The Pennsylvania State University, University Park,PA, USAb Noble Energy Corporation, Denver, CO, USAc Chevron Corporation, Houston, TX, USA

a r t i c l e i n f o

Article history:Received 23 April 2016Received in revised form22 August 2016Accepted 14 September 2016Available online 15 September 2016

Keywords:FaultingPermeabilityFault sealFriction

* Corresponding author.E-mail address: cjm38@psu.edu (C. Marone).

http://dx.doi.org/10.1016/j.marpetgeo.2016.09.0110264-8172/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

Faults in marine-sediment basins often dictate fluid flow and act as petroleum traps or seals. Perme-ability contrasts between these zones and the surrounding country rock can be large, depending on faultcomposition, stress, and strain history. Despite the importance of such faults, our understanding of theirfrictional properties and permeability is relatively poor. Here we report on a suite of laboratory exper-iments to assess the roles of fault composition, stress, and shear strain for dictating poromechanicalproperties of fault. Experiments were conducted at room temperature on synthetic fault gougecomposed of quartz, halite, and clay (illite shale or Ca-montmorillonite). We sheared layers that were 5e7 mm thick and measured fault permeability at effective normal stresses from 2 to 6 MPa for hydro-static conditions and after shear strains from 2 to 10. We find that fault permeability is highly sensitive toclay content, with permeabilities spanning 2e4 orders of magnitude under otherwise identical condi-tions. Permeability decreased up to 2 orders of magnitude with imposed shear strain >1 and ~1 order ofmagnitude with increasing normal stress from 2 to 6 MPa. During shear, halite deformed via both ductileflow and brittle failure, while quartz and clay particles formed force chains that spanned the layers anddefined shear localization fabrics. Samples with higher halite content exhibited greater permeabilityreduction, perhaps by plastic flow and pressure solution. Our results suggest that the permeability offaults in marine sediment is dictated by clay content, and that the most dramatic permeability changesoccur with relatively small fault throw (shear strain from 2 to 5). This highlights the importance ofunderstanding minor faults and fault sets when estimating subsurface fluid flow and potential reservoircompartmentalization.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Faults represent significant geologic features for subsurface fluidflow and are important in numerous transport processes withinsedimentary basins where they affect petroleum reservoir quality,hydrocarbon migration, and groundwater flow (Smith, 1966; Aydinand Johnson,1983; L�opez and Smith,1995; Evans et al., 1997; Aydin,2000; Kim et al., 2004; Baud et al., 2004). Fault permeability andporomechanical properties can impact deformation style in reser-voir rocks and a variety of tectonic environments (e.g., Arch andMaltman, 1990; Caine et al., 1996; Caine and Forster, 1999;

Crawford, 1998; Zhang et al., 1999; Fisher et al., 2001; Chuhanet al., 2003; Wibberley and Shimamoto, 2003; Odling et al., 2004;Mitchell and Faulkner, 2008, 2012; Ikari et al., 2009a,b; Balsamoand Storti, 2010; Faulkner et al., 2010; Foroozan et al., 2012;Carruthers et al., 2013; Gomila et al., 2016). In particular, faultscan act as fluid conduits in impermeable environments (e.g., shales,granites, basalts, etc.), fluid barriers in permeable environments(e.g., sandstones), and they may act to redirect flow in cases wherefaults develop significant permeability anisotropy (Knipe, 1992;Evans et al., 1997; Fisher and Knipe, 1998, 2001; Zhang and Tullis,1998; Bos and Spiers, 2000; Shipton et al., 2005; Benson et al.,2006; Fossen and Bale, 2007; Takahashi et al., 2007; Daniel andKaldi, 2008; Crawford et al., 2008; De Paola et al., 2009; Faulkneret al., 2010; Savage and Brodsky, 2011; Ballas et al., 2012; Ikariand Saffer, 2012; Faoro et al., 2012, 2013; Olierook et al., 2014;

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235 223

Scuderi et al., 2015; Seebeck et al., 2015).Fault permeability can be significantly lower than the wall rock,

especially in marine sediment environments where clay is inte-grated into fault zones (Caine et al., 1996; Evans et al., 1997; Gibson,1998; Crawford et al., 2008; Ikari et al., 2009a; Faulkner et al., 2010;Ishibashi et al., 2012; Mitchell and Faulkner, 2012; Harding andHuuse, 2015). In these environments, faults juxtapose and mixclays from shale alongwith sands and silts from other horizons (e.g,Bjùrlykke and Hùeg, 1997; Billi et al., 2003; Faulkner et al., 2010).Important processes include abrasional mixing, clay smearing, andpore throat clogging. Fault permeability is often dictated by poreconduit connectivity or narrow crack apertures, depending onmineralogy and development of shear fabric (Revil et al., 2002;Takahashi et al., 2007; Crawford et al., 2008; Ikari et al., 2009b;Niemeijer et al., 2010; Cheung et al., 2012; Scuderi et al., 2015).Indeed, small increases in clay concentrations can dramaticallyreduce fault permeability in sandy faults (Revil et al., 2002;Crawford et al., 2008; Faulkner et al., 2010). For example, Craw-ford et al. observed a four order of magnitude permeabilityreduction as clay content increased from 0 to 50% under otherwiseidentical conditions. While much work has been done to quantifythe roles of clay content, load, and strain on fault permeability (e.g.,Bentley and Barry, 1991; Antonellini and Aydin, 1994; Revil et al.,2002; Lothe et al., 2002; Wibberley and Shimamoto, 2003;Sternlof et al., 2004; Fortin et al., 2005; Crawford et al., 2008;Mondol et al., 2008; Tueckmantel et al., 2010; Ballas et al., 2012;Skurtveit, 2013; Scuderi et al., 2015), our understanding of therelationship between fault strength and permeability remainsrelatively poor.

Here, we focus on complex mixtures of synthetic fault gouge toilluminate the role of shear, fabric development and ductile flow,with application to faults found in association with salt domes andsurrounding country rock. One goal of this work is to provide inputfor geomechanical models of salt bodies, which require informationon both permeability and friction constitutive properties of thesalt-clay-sediment mixtures that often bound them.

Impermeable salt domes (Peach and Spiers, 1996) often haveassociated faults, and these features can act in concert to seal res-ervoirs (Jackson et al., 1994; Rowan et al., 1999; Hudec and Jackson,2007; Fort and Brun, 2012). In these cases, the fault zones oftenintegrate halite from the salt dome as well as nearby salt layers.Halite undergoes a range of deformation behaviors depending onthe chemistry, composition, materials it mixes with, stress, andstrain history on the fault, and thus its role for fault zone perme-ability is complex. Under fast slip conditions halite tends to deformbrittlely (e.g., Shimamoto, 1986; Bos et al., 2000a,b; Niemeijer et al.,2010; Wong and Baud, 2012) and should behave as a frameworkgrain. During slower deformation and for transient slip, halite likelyaccommodates shear via a combination of brittle and ductiledeformation and/or pressure solution (Rutter, 1983; Shimamoto,1986; Bos et al., 2000a,b; Niemeijer et al., 2008, 2010).

The purpose of this paper is to present results from a compre-hensive laboratory study of the permeability and frictional prop-erties of marine-sediment faults as a function of composition,stress, and shear strain. We studied synthetic faults composed ofquartz, halite, and illite shale or smectite clay. We report on fric-tional properties and fault-normal permeability as a function ofstrain and effective normal stress, focusing primarily on loweffective stress (<6 MPa) relevant for overpressured environments.

2. Methods and materials

We measured permeability and frictional properties of simu-lated fault zones as a function of normal stress and shear strain. Weconducted experiments in the double-direct shear configuration

under true triaxial stress conditions (e.g., Ikari et al., 2009a;Samuelson et al., 2009; Kaproth and Marone, 2014; Carpenteret al., 2015). The experimental geometry is shown in Fig. 1. Tosimulate faults in marine-sediment basins we conducted experi-ments on mixtures of quartz sand, illite shale or smectite clay (Ca-montmorillonite), and halite (Table 1 and Fig. 2). Our samplecompositions were designed to span the transition between clast-supported and matrix supported fault gouge, where fault zoneporosity is lowest and permeability changes are strongest (e.g.,Faulkner et al., 2010). We used commercial silt-sized quartzfoundry sand (F110) and Ca-montmorillonite (sieved to <125 mm).Illite shale and pure halite were ground in a rotary mill and sievedto <125 mm. The illite shale is primarily illite with some quartz,plagioclase, and minor kaolinite (Ikari et al., 2009a). Our experi-ments were carried out under conditions where pressure solutionand ductile deformation of halite is operative at low strain rates andduring holds, but halite and the other minerals deform via brittledeformation for the shearing rates used (e.g., shear displacementrate of 10 mm/s or shear strain rate of ~ 2 � 10�3 s�1). Our pore fluidwas a saturated NaCl brine, 35.7 g/ml, obtained by adding pure NaClto deionizedwater. We ensured NaCl saturation by regularly addingsalt grains to the brine bath.

We deformed synthetic faults as layers in the double directshear configuration under constant effective stress normal to thelayers (Fig. 2). In this configuration, two identical layers of fault-gouge are sandwiched between three forcing blocks (Fig. 1). Allstresses, displacements, and other measurements reported here aregiven for one layer of the double direct shear geometry. Sheardisplacement at the fault boundary is imposed by driving thecentral block with a hydraulic ram, inducing shear within the faultgouge layers. We arranged the true triaxial apparatus for constantconfining pressure, with normal stress and shear stress applied tothe sample directly via hydraulic pistons (Fig. 1). All stresses andfluid flow rates were maintained by fast-acting, servo-hydrauliccontrol. Strain gauge load cells, accurate to ±0.1 kPa, measuredapplied stresses. Direct current displacement transducers (DCDTs),accurate to ±0.1 mm, measured shear and normal displacement onthe faults. Pressure transducers, accurate to ±7 kPa, measured theconfining and pore pressure. Stresses, pressures, and displacementswere digitally recorded at 10 kHz with a 24 bit system, and wereaveraged to 10e100 Hz for storage.

Fault layers were constructed to specific initial thicknesses from5 to 7 mm. To ensure reproducibility of the tests we weighed thesample material used to construct each layer. Layer thickness wasmeasured carefully on the bench during sample construction andthen again in the testing machine once a small normal load wasapplied. These measurements were accurate to ±50 mm. Layerthickness was also measured at the end of the run, prior tounloading, which allowed us to check and verify the thickness andthickness change compared to that measured via the DCDT. Wecalculated average shear strain gwithin the layer as the sum of dx/h,where dx is a given shear displacement increment (normally <1 mmfor our digital sampling rates) and h is the measured layer thick-ness, which is updated continuously during shear.

Porous steel frits act as the interface between the steel forcingblocks and the sample layers, distributing the pore pressure andallowing linear flow across the sample (parallel to sn; Fig. 1C). Thesample assemblies were sealed with flexible latex jackets to isolatethe sample from the confining oil (Fig. 1D). Pore pressures wereapplied to the sample through the forcing blocks, with upstreampressure (PPA) applied through the center block to both samplelayers and downstream pressure (PPB) applied simultaneously tothe two side blocks of the double direct shear configuration(Kaproth and Marone, 2014).

Fig. 3 shows representative shear stress curves for three

Fig. 1. Deformation apparatus and experiment setup. Experiments were conducted in a (A) biaxial apparatus with a (B) true triaxial pressure vessel using the (C) double direct shearassembly. Confining pressure, PC, pore pressure inflow, PPA, and pore pressure outflow, PPB were applied through steel lines to the sample blocks. (C) Fault gouge was deformed inlayers. Normal stress was applied horizontally across both sample layers, and shear stress was applied vertically to the center block, inducing shear within the sample layers. Porefluid access was via channels and porous metal frits in the side blocks (D). The sample and pore pressures are isolated from the confining pressure with latex rubber jackets.

Table 1Permeability measurements for all experiments.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235224

Fig. 2. Ternary diagram showing the fault gouge compositions used in this study.Experiments were carried out on mixtures of quartz, halite, and clay (illite shale ormontmorillonite). A total of 18 experiments were conducted. Symbol color andexperiment number (e.g., p3979 is 100% Qtz) denote mixture composition here and insubsequent figures. (For interpretation of the references to colour in this figure legend,the reader is referred to the web version of this article.)

Fig. 3. Friction curves from three representative experiments. See Fig. 2 for compo-sition details. Each curve shows a stage of elastic loading followed by the onset ofplastic strain and shear deformation. After peak stress materials with high clay contentweaken dramatically, and for the sample with no quartz the coefficient of residualsliding friction (m) is ~0.3. Vertical line segments after 5 mm of shear displacementrepresent ‘hold’ periods designed to measure frictional restrengthening and perme-ability. Frictional strength increases with increasing quartz content; maximum residualfriction values for steady sliding are 0.6e0.7, which is typical for a clast-supportedmixture. In each experiment shear was stopped (hold periods) during permeabilitytests and then resumed, resulting in a drop in shear stress followed by a large peak thatrepresents sample lithification and frictional healing. Several hold tests were done inexperiment p3978.

Table 2Load steps for each experiment.

sn' sn PC Pp

6 4.5 3 2.44 3.0 2 1.62 1.5 1 0.80.5 0.6 0.35 0

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235 225

experiments. Upstream and downstream pore pressures (PPA andPPB, respectively) were used to drive saturated NaCl brine throughthe sample parallel to sn and perpendicular to the shear direction(Fig. 1).

We used a standard experimental protocol for all experiments(Table 1). The sample assemblies were loaded into the vessel and asmall initial normal load was applied (0.6 MPa sn). At this pointlayer thickness was measured under load with calipers, and addi-tional thickness changes were tracked with the DCDTs. The initiallayer thickness is known to±' 50 mm, and changes in layer thicknessare measured to ±0.1 mm. The pressure vessel was then filled withoil, and a small confining pressure was applied (0.35 MPa). Tosaturate the sample initially we removed air from the sample bypulling a vacuum from PPB. After the vacuum was drawn down tothe lowest point for ~1 min, pore fluid was introduced through PPAat a low constant flow rate. Flow was maintained until the effluentwas free of air for ~20 min (~40 min total). After initial saturation,sn’was increased to 2, 4, and 6 MPawith a permeability test at eachload (see Table 2 for descriptions of sn, pore pressure, and confiningpressure histories). For the double direct shear configuration in thetrue triaxial apparatus effective normal stress on the layers is givenas: sn0 ¼ sn e PP þ 0.5PC, where PP is the average pore pressure, PC isthe confining pressure, and the term 0.5PC represents the affect ofconfining pressure on normal load due to the sample geometry(e.g., Samuelson et al., 2009; Kaproth and Marone, 2014).

Fig. 4 shows the time history for our standard loading procedurefor a representative experiment. Fluid was driven through thesample by applying a constant upstream pressure PPA and a con-stant flow rate Q into the downstream fluid reservoir PPB (Fig. 1).Thus, flow occurs from the center block of the double direct shear(DDS) configuration to the side blocks where it is collected andmeasured as the total flow through both layers. We determinedpermeability following Darcy's Law, k ¼ Qnh/(A DPP), where Q is thevolumetric flow rate, n is the fluid viscosity, h is the layer thickness,(we use the actual value of h which is measured continuously with

the DCDT), A is the sample contact area (~0.005 m2), and DPP ¼ (PPAe PPB) is the pressure difference across the sample layers. All of ourmeasurements are referenced to one layer, including h, A, and Q.Fig. 5 shows the full details of a representative permeability test.Uncertainty in our permeability measurements (e.g., Fig. 5; Table 1)was determined from the measured flow rate and fluid pressure,accounting for specific storage effects as well as variations in flowrate and pressure.

Fig. 6 shows linear flow rates (q) and the differential pressuresduring each of the five permeability tests in p3911, where q ¼ Q/A.To limit dissolution, we generally used very low constant flow rates(q < 50 mm/s) and flow volumes, and pore fluid was typically flowedin and out of the sample (e.g., Fig. 4, sawtooth pattern). These stepslimited new pore water interaction with the sample. In some of theearly experiments, permeability tests were carried out under con-stant pressure conditions, which generally required larger pressuredifferences and flow rates, and may have caused minor halitedissolution, despite the fact that the pore fluid was a brine satu-rated with NaCl.

Fig. 4. Complete history for a representative experiment. Each line is noted with the initial and final values, for example shear stress starts at 0 and ends at 2.8 MPa, after 15,500 s.The y axis for each line is linear. Normal stress is applied initially to 0.6 MPa and then confining pressure is applied followed by pore pressure. Note history of effective normal stresssn’ shown below Pore Pressure A in the upstream reservoir (Fig. 1). After 1 h of initial fluid saturation, samples were loaded to 2 MPa sn’ through a combination of increased porepressure, sn, and confining pressure, PC. A permeability test (i.e. Perm. A) was conducted at constant flow rate after the sample reached steady-state compaction. Following the permtest, the sample was loaded to 4, and 6 MPa, and then the sample was sheared. Each of these steps preceded a permeability test.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235226

At 6 MPa sn0 each sample was sheared 10 mm, after which a k

test was conducted (Fig. 3). Initial shear was corrected for theelasticity of the latex jacket following standard procedure(Samuelson et al., 2009; Ikari et al., 2009a), accommodating ~ 5mmof the initial vertical displacement. Each sample was again sheared10mmand the final k test was conducted. Fig. 6 shows permeabilitytests at 2, 4, and 6 MPa sn

0 with zero strain, and at low and highstrain at 6 MPa sn0 (i.e. Perm. Ae E) for a representative experiment.Under each of these conditions and prior to each k test, the samplewas allowed to compact until compaction rates were less than0.04 mm/s (Fig. 4), limiting permeability transients.

Fig. 5. Representative permeability test with constant flow rate. During each k test aconstant flow rate was applied at PPB, driving a differential pressure between PPA andPPB. We measure permeability once flow reaches steady-state, with uncertainty givenby noise and small fluctuations in differential pressure.

3. Results

Our fault zones exhibited permeabilities from >2 � 10�14 to~3 � 10�19 m2, which is within the range of capability for ourtesting machine (Ikari et al., 2009a; Samuelson et al., 2009). Fig. 6bshows the permeability history for p3911, with permeabilitiesranging from ~4� 10�15 to ~8� 10�17 m2. Fig. 7 shows permeabilityas a function of load and strain for smectite experiments, and Fig. 8shows permeability for illite bearing experiments. The highestpermeabilities occurred for pure quartz sand at low stresses,exceeding the capabilities of our system in some cases.

In general, permeability decreased as a function of decreasingquartz content and increasing matrix content (e.g., clay and halite).Fig. 9 shows permeability contours interpolated between a numberof our experiments, prior to and after the application of shearstrain. From 75% to 33% quartz content the permeability decreasedby ~1e2 orders of magnitudewith shear strain. In both the smectiteand illite samples, the lowest permeabilities occur for mixtures of

33/54/12% by mass quartz/clay/halite, closely followed by the 33/33/33% mixtures. Superimposed on these values are trends ofpermeability decrease with load and strain (Fig. 9).

We found that permeability decreased as a function of

Fig. 6. Details of each permeability test from one experiment, p3911 (See Fig. 4). Panel A shows the linear flow rate, q ¼ Q/A, and resultant differential pore pressure for each k test.With increasing load and shear strain, we decreased q to limit DP, as k tended to decrease with each step. Low flow rates were also used to limit halite dissolution and momentumtransfer effects between fluid and grains. Panel B shows the change in permeability as a function of load (left side) and shear strain (right side), highlighting the tendency for k todecrease with load and strain.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235 227

increasing normal stress and shear strain (Figs. 6e8). Withincreasing normal load from 2 to 6 MPa, all materials experienced apermeability reduction of 0e1 orders of magnitude. Similarly, theapplication of shear strain (from 0 to 2e5 g) caused a permeabilityreduction of 0e2 orders of magnitude. At higher strains however(5e10 g), some samples exhibited permeability increases of up to 1order of magnitude (Figs. 7e8).

Fig. 10 shows permeability change with load and strain as afunction of quartz content and matrix composition (clay: haliteratio). With increasing effective stress, the observed drop inpermeability was generally insensitive to quartz content, matrixcomposition (clay/halite ranging from 5/95% to 81/19%), and claytype. With initial shear strain, smectite experiments experienced

Fig. 7. Permeability as a function of load (left side) and strain (right side) across for montmpermeabilities spanned ~2 orders of magnitude, OM. For 100% quartz, p3979, we measuredexperiments k decreased 0e1 OM with increasing load. Similarly, k decreased by 0e2 OM wiAt higher strains however, k tended to level out or increase slightly, as was typically the ca

more k reduction than illite (Fig. 10B0 vs. 10B, respectively), and thisk reduction was enhanced by decreasing quartz content. Forexample, permeability decreased ~ 1e2 orders of magnitude with<50% quartz content and only ~0e1 orders of magnitude for higherquartz contents (Fig. 10B0).

With the progression from low to high strain, permeabilitytended to increase. The greatest k increase occurred for mixtureswith high matrix content. Smectite bearing mixtures exhibited thegreatest k increase from low to high strain (Fig. 10C0), perhaps inresponse to the large k decrease that occurred with initial strain(Fig. 10B0).

To study the role of quartz in determining permeability changes,we compared six experiments with varied quartz content but equal

orillonite experiments. In general, k increased with high quartz content, and baselinevery high k, exceeding the system measurement limit (k ~ 2 � 10�14 m2). Across all

th increasing strain and the reduction was higher for samples with higher clay content.se for high matrix content experiments at very high strains (g > 8).

Fig. 8. Permeability as a function of load and strain for illite shale experiments. Permeability increased with increasing quartz content, with baseline permeabilities spanning ~2.5OM. k decreased with increasing load by 0e1 OM. k decreased by 0e2 OM as shear strain increased from 0 to ~3; the effect was greater for the samples with the highest clay content.k did not change dramatically at higher shear strains, and tended to increase in matrix-rich experiments at higher strains (g > 6).

Fig. 9. Permeability as a function of mixture composition with contours drawn via interpolation between most experiments. Illite clay samples on left panels, smectite clay on rightpanels. All data are for 6 MPa effective normal stress. Panels A and A0 show data for strain ¼ 0. Panels B and B0 show data for low strain (see Figs. 7 and 8). Permeability decreasedwith increasing clay content. k was lower for samples with halite relative to quartz at constant clay content. Note that permeability does not show systematic differences betweenthe illite and smectite clay experiments.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235228

Fig. 10. Trends of permeability loss as a function of load (A and A0), low strain (B and B0), and high strain (C and C0) for illite and smectite samples, respectively. Colors denote matrixcomposition (See Fig. 1). With increased load, k decreased by 0e1 orders of magnitude, with little dependence on clay type (A vs. A0), quartz content, or matrix composition. At lowstrain, permeability decreased by 0e2 OM, with the strongest change occurring for clay-rich experiments. At higher strain, k changed little, with the strongest increases occurringfor experiments with the highest final strains (see Table 1) and clay contents. (For interpretation of the references to colour in this figure legend, the reader is referred to the webversion of this article.)

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235 229

parts clay and halite (Fig. 11 inset). Fig. 11B and C highlight k as afunction of quartz content at different loads and strains, respec-tively. At the highest loads (6 MPa sn0), permeability was lowest for33% quartz (~5 � 10�17 m2), and highest for 100% quartz(>2 � 10�14 m2). The fault zone with no quartz exhibited slightlyhigher permeability than that with 33% quartz. Across these ex-periments, k decreased by 0e1 orders of magnitude with increasedquartz content (Fig. 11B).

With shear strain, similar to the affect of load, permeability waslowest for the mixture with 33% quartz, 33% smectite, and 33%halite by mass, and generally increased with quartz content(Fig. 11C). As identified previously, the lowest permeabilities formost of these mixtures occurred after initial shear strain (2e4 g),followed by slight k increase with further strain (Fig. 11C).

As with permeability, fault friction is strongly dependent onquartz content. Fault layers dominated by quartz, >57%, exhibitsimilar frictional strengths, at or above 0.6 coefficient of friction(m ¼ t/sn0) (Fig. 11A). With higher matrix content (�33% quartz),friction never exceeded 0.6 and the fault gouge continuouslyweakens with strain following an initial peak. Weakening behav-iors such as these indicate matrix-dominated sliding, perhaps withfabric formation and clay-particle alignment (Rutter et al., 1986;Logan et al., 1992; Haines et al., 2013).

We used microstructural images to document the evolution ofshear fabric with shear strain (Figs. 12e14). Samples were collectedfor thin sections following the final permeability test, and thusthese images document the highest strains attained. Halite parti-cles appear as bright-white, quartz particles are large and gray, andthe graymatrix is smectite clay (Fig.12). For high quartz content the

shear zones are framework grain supported (Fig. 12AeC), but theyare matrix supported for lower quartz content (Fig. 12DeE). Fabricgeneration appears to occur primarily in the matrix-supportedgouges. Fig. 13 shows higher magnification views of the micro-graphs from Fig. 12. Experiments with higher quartz contents donot develop a strong matrix fabric (Fig. 13AeC, C0), but fabric is welldeveloped with high matrix content (Fig. 13DeE). Most pore space,especially along quartz grain boundaries, is open as a result ofsmectite dehydration and densification following each experiment.However, some of this pore space was likely openwhile the samplewas hydrated and intact. In many of these experiment (e.g., p3976)halite grains appear to be in brittle contact with other halite orquartz grains. This allows for accumulation of clay particles be-tween halite grains, potentially leaving pore-space between quartzgrains. Halite pressure solution welding is evident throughoutthese micrographs and in some cases the halite appears to flowwith the clay fabric (Fig. 13). For smectite dominated layers, weobserve that halite welds with other grains and is in brittle contactwith some quartz grains (Fig. 14). In these cases, P-shears devel-oped throughout the layer and halite grains are seen flowing withinthe clay fabric (Fig. 14).

We conducted permeability tests after the sample had com-pacted. As such, halite may have acted as a matrix material duringthese tests, infilling pore space via pressure solution. In some cases,halite grains mix with clay particles to form localized shear zones.Micrographs also indicate halite grains in stress shadows of clastnetworks and force chains (Fig. 13). This indicates that halite is notsimply a ‘matrix’ material that flows to fill pore space, but rather itcan remain in stress shadows and deform brittlely, both of which

Fig. 11. Fault properties as a function of quartz content, with constant clay to haliteratio (see inset to panel A). Sample strength evolves from low to high as quartz contentincreases (A). In particular, matrix dominated materials weaken dramatically withshear strain, likely resulting from fabric development (Haines et al., 2013), and quartzdominated materials maintain friction values above 0.6. Quartz dominated materialswith some matrix exhibit more healing, likely resulting from the halite content, andgenerally higher stresses, resulting from pore-space infilling (Kaproth et al., 2010).Panels B and C show the permeability evolution with quartz content during theseexperiments as a function of load and strain, respectively. In general, increased quartzcontent results in increased k, with an apparently smooth trend from 100% to 33%quartz. Permeability appears to increase slightly from 33% to 0% quartz. Experimentp3976 is slightly off trend with the other experiments, but after small strain comes intoline with other experiments, indicating the effects of shear fabric development duringthe run-in period.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235230

tend to result in higher pore space than suggested by some previ-ous models (e.g., Revil et al., 2002).

4. Discussion

We find that permeability and frictional strength are stronglysensitive to clay content and shear strain. Faults with small con-centrations of clay have dramatically lower k values than those thatdo not (Fig. 11). Overall our permeability values range from aminimum of 5 � 10�19 m2 to the upper bound that we can measureaccurately, which is 2 � 10�14 m2. This included a range of strainsand loads for quartz/clay/halite end-member mixtures of 100/0/0%,33/54/12%, and 5/5/90%, respectively. These values are consistentwith previously reported values for pure quartz, illite,montmorillonite-smectite, halite, and mixed granular materials(Peach and Spiers, 1996; Zhang and Tullis, 1998; Crawford et al.,2008; Ikari et al., 2009a; Niemeijer et al., 2008, 2010). Forexample, Ikari et al., 2009a,b reported 2 � 10�16 m2 permeability

for a 50% quartz, 50% montmorillonite mixture (sn0 ¼ 12 MPa),compared to our value of 8 � 10�17 m2 for a similar mixture.

With increased effective stress from 2 to 6 MPa, we observe upto 1 order of magnitude permeability reduction across all of ourmaterials (Figs. 7, 8 and 10). Previous studies show similar magni-tudes of k reduction, especially at such low normal loads, whereanelastic compaction is dominant (Faulkner and Rutter, 2003;Crawford et al., 2008; Ikari et al., 2009a).

With shear, fault permeability decreased by up to 2 orders ofmagnitude and tended to be enhanced in clay-rich materials(Figs. 7, 8, 10 and 11). These changes are in general agreement withprevious laboratory works on synthetic fault zones (Zhang andTullis, 1998; Crawford et al., 2008; Ikari et al., 2009a). Sheardriven compaction, fabric formation, and abrasional mixing all tendto assist k reduction with fault development (Faulkner et al., 2010).Abrasional mixing is especially important in nature because it al-lows commingling of different wall rock materials, mixing clay inwith clasts. Similar processes likely occur during our experiments,helping to mix the gouge and modify the shear fabric (e.g., Rathbunand Marone, 2010). These processes are likely the cause of what wesee for experiment p3976, which deviates from other runs withload but falls on trend following shear (Fig. 11).

Shear fabric formation is particularly evident as a function ofshear strain for the clay rich fault zones (Figs. 12e14). Frictionvalues for higher quartz content fault zones indicate frictionalsliding amongst clasts. In particular, frictional sliding with only 57%quartz indicates that halite is participating in clast-type deforma-tion during shear, as brittle deformation of halite may be operativeat these load (6 MPa sn

0) and strain-rate (10 mm/s) conditions(Shimamoto, 1986). The highest frictional strengths occur withsome matrix content (e.g., 57e75% quartz), indicating that materialin the pore space can obstruct efficient frictional deformation(Kaproth et al., 2010). Peak stress values following holds are likelylarge resulting from frictional healing via pressure solution of halite(e.g., Marone, 1998; Bos et al., 2000b; Niemeijer et al., 2008, 2010;Carpenter et al., 2016). These peaks may also indicate that theframework clasts are in direct contact with one another, since clayveneer tends to limit fault healing (Dewers and Ortoleva, 1991; Boset al., 2000a).

Halite appears to have multiple roles in affecting the perme-ability and mechanical properties of the fault zones. It deforms viaductile deformation and pressure-solution (e.g., Bos et al., 2000b)and also via cracking and other brittle mechanisms. Halite grainsappear to participate in force chains (e.g., Fig. 13C0) that carry stressin clast-supported materials during shear (Daniels and Hayman,2008), suggesting brittle failure. However, halite also shows char-acteristics of ductile flow, as indicated by grain textures andboundary welding, which is presumably a result of pressure solu-tion (e.g., Fig. 13E) (Rutter, 1983; Bos et al., 2000b; Niemeijer et al.,2008).

Throughout our experiments halite particles show characteris-tics of both brittle and ductile deformation. Some halite grainsappear to participate in force chains, which carry stress in a clast-supported gouge (e.g., Daniels and Hayman, 2008) and showsome evidence of grain cracking. However, halite grains also showevidence of ductile deformation, such as bending around quartzand halite clasts, and pressure solution. In particular, Fig. 14 showshalite grains welded to quartz particles and clay fabric, evidence forpressure solution (e.g., Bos et al., 2000b). We suspect that somehalite grains deform brittlely during shear, resulting in clast sup-ported frictional behavior with low quartz concentrations (e.g., 33%quartz; Fig. 11). However, ductile deformation and pressure solu-tion likely dominate halite deformation during holds. Similarductile processes may occur for halite grains sitting in stressshadows during shear.

Fig. 12. SEM micrographs from five experiments with varied quartz content and constant halite to clay ratio (e.g., Fig. 11). Each of these thin sections was made following the finalpermeability test, and are thus at high strain (Table 1). Halite particles are bright-white, quartz particles are large and gray, and the gray matrix is smectite clay (individual particlesare not visible, at <1 mm diameter). At high quartz content the faults are framework grain supported (AeC), but are matrix supported at lower quartz content (DeE). Fabricgeneration appears to only occur in the matrix-supported gouges. Although salt acts as matrix during holds, during shear it appears to be load-bearing, acting as part of the grain-framework skeleton. In particular, p3910 shows halite grains apparently within force chains with quartz grains. Since these thin sections were made following the final hold, saltgrains show strong evidence plastic and pressure solution behavior, which are highlighted in Fig. 13.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235 231

Although halite acts as matrix during permeability tests, it ap-pears that relicts of shear-driven brittle deformation are main-tained during holds and k tests. The parametric analysis (Fig. 11)indicates that k was lowest for the mixture with 33% quartz, 33%halite, and 33% smectite by mass, or 28% quartz and 72% matrix byvolume, whereas previous models suggest that k should be lowestwith 25e40 vol % matrix (Revil et al., 2002; Crawford et al., 2008).Additionally, SEMmicrographs show halite grains involved in forcechains (Fig. 13). These observations indicate that halite does notsimply participate as matrix. This suggests that minimum porosityand perhaps minimum k occurs with some higher clay content,

making up for the non-ideal distribution of halite grains.In experiments with small amounts of quartz (e.g., p3909), the

material behaves as if it were matrix dominated regardless of halitecontent. Fig. 13D shows that p3909 develops fabric within thematrix, with clay and halite aligning in the P orientation. Thismaterial also weakened significantly during shear (Fig. 11A), high-lighting the role of shear fabric generation. In this case, fluid flow isprimarily through the matrix where it is partially retarded byimpermeable quartz grains. We suspect that kmight be lower witha slightly higher clast concentration, at the threshold between amatrix and clast supported gouge (Revil et al., 2002). Additionally,

Fig. 13. Higher magnification SEM micrographs from Fig. 12. High quartz experiments do not develop a strong matrix fabric (A-C, C0), but fabric is well developed with high matrixcontent (DeE). Most pore space, especially along quartz grain boundaries, is open as a result of smectite dehydration and densification following each experiment. However some ofthis pore space was likely open while the sample was hydrated and intact. In many of these experiment (e.g., p3976) halite grains appear to be in brittle contact with other halite orquartz grains, indicated by B. This allows for accumulation of clay particles between halite grains, potentially leaving pore-space between quartz grains, indicated by f. Halitewelding (via pressure solution) is evident throughout these micrographs, annotated asW, and in some cases the halite appears to flow with the clay fabric, noted as P. Panels C and Fhighlight sister experiments with similar mixture ratios, but contain smectite (montmorillonite) clay and illite shale, respectively. These mixtures are similar, but the matrix appearsmore granular with illite than smectite, owing to the contrast in particle size and some quartz content of the illite shale.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235232

we expect that for a given concentration of clay k will be prefer-entially lower with halite over quartz because halite can infill someneighboring pore space via pressure solution (Dewers and Ortoleva,1991; Niemeijer et al., 2008) and ductile deformation. This may beespecially true during times of low shear strain, when halite canflow to occupy porosity between quartz grains and clay fabric. Theeffect of pressure solution may also be enhanced during holds bythe presence of clay particles, which tends to increase diffusionrates (Dewers and Ortoleva, 1991; Renard et al., 2001).

4.1. Application in reservoir settings and implications for faults

While the type of mixed-mode deformation that occurs in ourlaboratory samples may be relevant for faults that slip seismically

or for faults with low sn0 (Shimamoto, 1986; Davison, 2009), the

application of these data to faults in reservoir rocks is less clear.Several factors may promote aseismic fault creep, such as slowerloading rates, higher effective stresses, or higher temperatures.Under these conditions halite may accommodate fault slip domi-nantly via ductile deformation and pressure solution. Withoutbrittle deformation, halite would behave as a matrix material and,based on previous work, minimum porosity should occur from 60to 75 vol % quartz (Revil et al., 2002). In this case permeabilitywould likely be lowest for near-pure mixtures of halite and clay,since these mixtures develop mylonitic fabric (Bos et al., 2000a)and halite grains are nearly impermeable (Peach and Spiers, 1996).

In our matrix supported fault zone materials, non-myloniticfabric is generated within the matrix material, which may have

Fig. 14. SEM micrograph from smectite dominated experiment, p3821 (54% smectite,33% quartz, 12% halite). Like the other experiments, halite welds with other grains (W)and is in brittle contact with some quartz grains (B). P-shears developed throughoutthis material, typical of clay-rich materials. Additionally, halite grains flow with andweld to the clay fabric, and apparently do not locally disrupt the clay fabric, unlikequartz grains.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235 233

implications for fault permeability decrease at low strain and in-crease at higher strains. Fabric generation appears to be enhancedwith greater clay content (Fig. 13D vs. Fig. 13B), and generally alignsclay particles into the P-orientation or the fault parallel Y or Borientation (e.g., Haines et al., 2013). Fault parallel clays may retardacross-fault flow, the direction tested here, perhaps reducingpermeability by an order of magnitude in that direction (Faulkneret al., 2010). With greater shear however, macro-fabric features(e.g., Riedel, P, and Y shears) are developed (Fig. 14) (Logan et al.,1992; Haines et al., 2009, 2013), and fluid flow may be enhancedalong these features (Arch and Maltman, 1990; Zhang and Tullis,1998; Zhang et al., 1999). This may be especially true along highangle P-shears, which readily develop in high clay content experi-ments (Fig. 14). Thus k may increase in clay rich fault gouge fromlow to high strain (Fig. 10), similar to observations by Crawfordet al., (2008). Previous studies at higher loads show continued kdecrease with strain, likely owing to the effects of comminutionand compaction (Ikari et al., 2009a; Faulkner et al., 2010), whichwas limited in our low-stress, quartz-rich system (e.g., note theintact grains visible after shear in Fig. 13). Finally, we note that ourexperiments, with halite saturated brine, apply to faults in halitebearing units where fluid flow rates are slow compared to the rateof dissolution. Moreover, the observation that permeability re-sponses are similar between the illite and smectite bearing lithol-ogies suggests that halite may prevent swelling of the smectitegrains.

5. Conclusions

Our experiments were designed to investigate the permeabilityevolution of marine sediment faults with intermixed halite andclay. We found that fault permeability has strong dependence onclay content, reducing k by up to two orders of magnitude underotherwise identical conditions. For permeability we found thathalite and quartz are generally interchangeable, with slightly

enhanced k reduction with halite resulting from enhancedcompaction during holds. We studied loading conditions analogousto those of the seismic cycle, with periods of relatively fast slipfollowed by periods of quiescence. Halite grains show evidence forboth brittle and ductile deformation, behaving brittlely duringshear and ductily during hold periods. We found that halite cannotbe consideredmatrix whenmakingminimum porosity estimates ofa mixture; minimum porosity will occur instead with a greaterproportion of fine particles.

For all compositions of marine sediment faults tested here wefound that permeability was greatly reduced with small shearstrain (g < 5), up to two orders of magnitude. Greater strain hadlimited effect on permeability, and occasionally enhanced perme-ability, presumably due to disruption of flow retarding shear fabric.Similarly, we found that fault permeability decreased by<1 order ofmagnitude with increased load (<6 MPa sn

0). These results are ingeneral agreement with previous works showing the permeabilityeffect of compaction with load and abrasional mixing and claysmearing with shear (Zhang and Tullis, 1998; Faulkner and Rutter,2001, 2003; Crawford et al., 2008; Ikari et al., 2009a). Our resultsindicate that permeability of faults in marine sediment is dictatedstrongly by clay content, and that the most dramatic permeabilitychanges occur with relatively small fault throw. These findingssuggest that minor faults, which may be difficult to detect inseismic data, may have dramatic implications for reservoircharacterization.

Acknowledgements

We thank S. Swavely for help in the laboratory, D. Elsworth forhelpful discussions, and A. Billi and an anonymous reviewer forthorough reviews and insightful comments that improved themanuscript. This work is the result of support from the ChevronEnergy Technology Company. This support is gratefully acknowl-edged. All data used in this article are available by contacting thecorresponding author.

References

Antonellini, M., Aydin, A., 1994. Effect of faulting on fluid flow in porous sandstones:petrophysical properties. Am. Assoc. Pet. Geol. Bull. 78, 335e377.

Arch, J., Maltman, A., 1990. Anisotropic permeability and tormosity in deformed wetsediments. J. Geophys. Res. Solid 95 (89), 9035e9045.

Aydin, A., 2000. Fractures, faults, and hydrocarbon entrapment, migration and flow.Mar. Pet. Geol. 17, 797e814.

Aydin, A., Johnson, A.M., 1983. Analysis of faulting in porous sandstones. J. Struct.Geol. 5.

Ballas, G., Soliva, R., Sizun, J.P., Benedict, A., Cavailhes, T., Raynaud, S., 2012. Theimportance of the degree of cataclasis in shear bands for fluid flow in poroussandstone, Provence, France. AAPG Bull. 96 (11), 2167e2186.

Balsamo, F., Storti, F., 2010. Grain size and permeability evolution of soft-sedimentextensional sub-seismic and seismic fault zones in high-porosity sedimentsfrom the Crotone basin, southern Apennines, Italy. Mar. Pet. Geol. 27 (4),822e837.

Baud, P., Klein, E., Wong, T.F., 2004. Compaction localization in porous sandstones:spatial evolution of damage and acoustic emission activity. J. Struct. Geol. 26 (4),603e624.

Benson, P., Schubnel, A., Vinciguerra, S., Trovato, C., Meredith, P., Young, R.P., 2006.Modeling the permeability evolution of microcracked rocks from elastic wavevelocity inversion at elevated isostatic pressure. J. Geophys. Res. 111.

Bentley, M.R., Barry, J.J., 1991. Representation of Fault Sealing in a Reservoir Simu-lation: Cormorant Block IV. Society of Petroleum Engineers, UK North Sea. Paper22667.

Billi, A., Salvini, F., Storti, F., 2003. The damage zone-fault core transition in car-bonate rocks: implications for fault growth, structure and permeability.J. Struct. Geol. 25, 1779e1794.

Bjùrlykke, K., Hùeg, K., 1997. Effects of burial diagenesis on stress, compaction andfluid flow in sedimentary basins. Mar. Pet. Geol. 14, 267e276.

Bos, B., Spiers, C., 2000. Effect of phyllosilicates on fluid-assisted healing of gouge-bearing faults. Earth Planet. Sci. Lett. 184 (1), 199e210.

Bos, B., Peach, C., Spiers, C., 2000a. Frictional-viscous flow of simulated fault gougecaused by the combined effects of phyllosilicates and pressure solution. Tec-tonophysics 327 (3e4), 173e194.

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235234

Bos, B., Peach, C., Spiers, C., 2000b. Slip behavior of simulated gouge-bearing faultsunder conditions favoring pressure solution. J. Geophys. Res. 105 (B7), 16699.

Caine, J.S., Forster, C.B., 1999. Fault zone architecture and fluid flow: insights fromfield data and numerical modeling. In: Haneberg, W.C., et al. (Eds.), Faults andSubsurface Flow in Shallow Crust: American Geophysical Monograph, vol. 113,pp. 101e127.

Caine, J.S., Evans, J.P., Forster, C.B., 1996. Fault zone architecture and permeabilitystructure. Geology 24, 1025e1028.

Carpenter, B.M., Saffer, D.M., Marone, C., 2015. Frictional properties of the active SanAndreas Fault at SAFOD: implications for fault strength and slip behavior.J. Geophys. Res. 120, 5273e5289. http://dx.doi.org/10.1002/2015JB011963.

Carpenter, B.M., Ikari, M.J., Marone, C., 2016. Laboratory observations of time-dependent frictional strengthening and stress relaxation in natural and syn-thetic fault gouges. J. Geophys. Res. 121, 1183e1201. http://dx.doi.org/10.1002/2015JB012136.

Carruthers, D., Cartwright, J., Jackson, M.P.A., Schutjens, P., 2013. Origin and timingof layer-bound radial faulting around North Sea salt stocks: new insights intothe evolving stress state around rising diapirs. Mar. Pet. Geol. 48, 130e148.

Cheung, C.S., Baud, P., Wong, T. f., 2012. Effect of grain size distribution on thedevelopment of compaction localization in porous sandstone. Geophys. Res.Lett. 39 (21).

Chuhan, F.A., Kjeldstad, A., Bjorlykke, K., Hoeg, K., 2003. Experimental compressionof loose sands: relevance to porosity reduction during burial in sedimentarybasins. Can. Geotech. J. 40 (5), 995e1011.

Crawford, B.R., 1998. Experimental fault sealing: shear band permeability de-pendency on cataclastic fault gouge characteristics. In: Coward, M.P.,Daltaban, T.S., Johnson, H. (Eds.), Structural Geology in Reservoir Characterisa-tion, vol. 127, pp. 27e48. Special Publication of the Geological Society of London.

Crawford, B.R., Faulkner, D.R., Rutter, E.H., 2008. Strength, porosity, and perme-ability development during hydrostatic and shear loading of synthetic quartz-clay fault gouge. J. Geophys. Res. 113.

Daniel, R.F., Kaldi, J.G., 2008. Evaluating seal capacity of caprocks and intraforma-tional barriers for the sequestration of CO2. In: PESA Eastern Australian BasinsSymposium III, pp. 475e484.

Daniels, K.E., Hayman, N.W., 2008. Force chains in seismogenic faults visualizedwith photoelastic granular shear experiments. J. Geophys. Res. 113.

Davison, I., 2009. Faulting and fluid flow through salt. J. Geol. Soc. 166, 205e216.http://dx.doi.org/10.1144/0016-76492008-064.

De Paola, N., Faulkner, D.R., Collettini, C., 2009. Brittle versus ductile deformation asthe main control on the transport properties of low-porosity anhydrite rocks.J. Geophys. Res. 114.

Dewers, T., Ortoleva, P., 1991. Influences of clay minerals on sand- stone cementa-tion and pressure solution. Geology 19, 1045e1048.

Evans, J.P., Forster, C.B., Goddard, J.V., 1997. Permeability of fault-related rocks, andimplications for hydraulic structure of fault zones. J. Struct. Geol. 19 (11),1393e1404.

Faoro, I., Elsworth, D., Marone, C., 2012. Permeability evolution during dynamicstressing of dual permeability media. J. Geophys. Res. 117. No. B1, B01310.

Faoro, I., Vinciguerra, S., Marone, C., Elsworth, D., Schubnel, A., 2013. Linkingpermeability to crack density evolution in thermally stressed rocks under cyclicloading. Geophys. Res. Lett. 40 http://dx.doi.org/10.1002/grl.50436.

Faulkner, D.R., Rutter, E., 2001. Can the maintenance of overpressured fluids in largestrike- slip fault zones explain their apparent weakness. Geology 29, 503e506.

Faulkner, D.R., Rutter, 2003. The effect of temperature, the nature of the pore fluid,and subyield differential stress on the permeability of phyllosilicate-rich faultgouge. J. Geophys. Res. 108.

Faulkner, D., Jackson, C., Lunn, R., Schlische, R., Shipton, Z., Wibberley, C.,Withjack, M., 2010. A review of recent developments concerning the structure,mechanics and fluid flow properties of fault zones. J. Struct. Geol. 32 (11),1557e1575.

Fisher, Q.J., Harris, S.D., McAllister, E., Knipe, R.J., Bolton, A.J., 2001. Hydrocarbonflow across sealing faults: theoretical constraints. Mar. Pet. Geol. 18, 251e257.

Fisher, Q.J., Knipe, R.J., 1998. Fault sealing processes in siliciclastic sediments. In:Geological Society, London, Special Publications, vol. 147, p. 117e134.

Fisher, Q.J., Knipe, R.J., 2001. The permeability of faults within siliciclastic petroleumreservoirs of the North Sea and Norwegian Continental Shelf. Mar. Pet. Geol. 18,1063e1081.

Foroozan, R., Elsworth, D., Flemings, P., Bilotti, F., Muhuri, S., 2012. The role ofpermeability in fault zones on the structural and hydro-mechanical character-istics of shortening basins. Mar. Pet. Geol. 29, 143e152.

Fort, X., Brun, J.-P., 2012. Kinematics of regional salt flow in the northern Gulf ofMexico. In: Alsop, G.I., Archer, S.G., Hartley, A.J., Grant, N.T., Hodgkinson, R.(Eds.), Salt Tectonics, Sedimentation and Prospectivity, vol. 363. GeologicalSociety, London, p. 265e287. Special Publications.

Fortin, J., Schubnel, A., Gueguen, Y., 2005. Elastic wave velocities and permeabilityevolution during compaction of Bleurswiller sandstone. Int. J. Rock Mech. Min.Sci. 42 (7e8), 873e889.

Fossen, H., Bale, A., 2007. Deformation bands and their influence on fluid flow.AAPG Bull. 91 (12), 1685e1700.

Gibson, R.G., 1998. Physical character and fluid flow properties of sandstone-derived fault zones. In: Coward, M.P., Daltaban, T.S., Johnson, H. (Eds.), Struc-tural Geology in Reservoir Characterisation, vol. 127, pp. 83e98. Special Publi-cation of the Geological Society of London.

Gomila, R., Mitchell, T.M., Arancibia, G., Cembrano, J.M., Faulkner, D.R., 2016.Palaeopermeability structure within fault zones: a snap-shot from

microfracture analyses in a strike-slip system. J. Struct. Geol. 83, 103e120.Haines, S.H., van der Pluijm, B. a., Ikari, M.J., Saffer, D.M., Marone, C., 2009. Clay

fabric intensity in natural and artificial fault gouges: implications for brittlefault zone processes and sedimentary basin clay fabric evolution. J. Geophys.Res. 114.

Haines, S.H., Kaproth, B.M., Marone, C., Saffer, D., van der Pluijm, B. a., 2013. Shearzones in clay-rich fault gouge: a laboratory study of fabric element develop-ment and evolution. J. Struct. Geol. 114 (B5).

Harding, R., Huuse, M., 2015. Salt on the move: multi stage evolution of salt diapirsin The Netherlands North Sea. Mar. Pet. Geol. 61, 39e55.

Hudec, M.R., Jackson, M.P.A., 2007. Terra infirma: understanding salt tectonics. EarthSci. Rev. 82, 1e28.

Ikari, M.J., Saffer, D.M., 2012. Permeability contrasts between sheared and normallyconsolidated sediments in the Nankai accretionary prism. Mar. Geol. 295e298,1e13. http://dx.doi.org/10.1016/j.margeo.2011.11.006.

Ikari, M., Saffer, D., Marone, C., 2009a. Frictional and hydrologic properties of clay-rich fault gouge. J. Geophys. Res. 114 (B5), B05409.

Ikari, M.J., Saffer, D.M., Marone, C., 2009b. Frictional and hydrologic properties of amajor splay fault system, Nankai subduction zone. Geophys. Res. Lett. 36 (20),L20313.

Ishibashi, T., McGuire, T., Tsuchiya, N., Elsworth, D., 2012. Permeability evolution ofcarbonate fractures: competition between stress and reactivity. Water Resour.Res. 48 (7).

Jackson, M.P.A., Vendeville, B.C., Schultz-Ela, D.D., 1994. Structural dynamics of saltsystems. Annu. Rev. Earth Planet. Sci. 22, 93e117.

Kaproth, B.M., Cashman, S.M., Marone, C., 2010. Deformation band formation andstrength evolution in unlithified sand: the role of grain breakage. J. Geophys.Res. 115, B12103.

Kaproth, B.M., Marone, C., 2014. Evolution of elastic wave speed during shear-induced damage and healing within laboratory fault zones. J. Geophys. Res. 119.

Kim, Y.-S., Peacock, D.C.P., Sanderson, D.J., 2004. Fault damage zones. J. Struct. Geol.26, 503e517.

Knipe, R.J., 1992. Faulting processes and fault seal. In: Larsen, R.M., Brekke, B.T., .Talleras, E. (Eds.), Structural and Tectonic Modelling and its Application to Pe-troleum Geology, vol. 1, pp. 325e342. NPF Special Publication.

Logan, J., Dengo, C., Higgs, N., Wang, Z.-Z., 1992. Fabrics of experimental fault zones:their development and relationship to mechanical behavior. In: Fault Mechanicsand Transport Properties of Rocks. Academic Press, pp. 33e67.

L�opez, D., Smith, L., 1995. Fluid flow in fault zones: analysis of the interplay ofconvective circulation and topographically driven groundwater flow. WaterResour. Res. 31 (6), 1489e1503.

Lothe, A.E., Gabrielsen, R.H., Hagen, N.B., Larsen, B.T., 2002. An experimental studyof the texture of deformation bands: effects on the porosity and permeability ofsandstones. Pet. Geosci. 8 (3), 195e207.

Marone, C., 1998. Laboratory-derived friction laws and their application to seismicfaulting. Ann. Rev. Earth Plan. Sci. 26, 643e696.

Mitchell, T.M., Faulkner, D.R., 2008. Experimental measurements of permeabilityevolution during triaxial compression of initially intact crystalline rocks andimplications for fluid flow in fault zones. J. Geophys. Res. 113.

Mitchell, T.M., Faulkner, D.R., 2012. Towards quantifying the matrix permeability offault damage zones in low porosity rocks. Earth Planet. Sci. Lett. 339e340,24e31.

Mondol, N.H., Bjorlykke, K., Jahren, J., 2008. Experimental compaction of clays:relationship between permeability and petrophysical properties in mudstones.Pet. Geosci. 14 (4), 319e337.

Niemeijer, A., Marone, C., Elsworth, D., 2008. Healing of simulated fault gougesaided by pressure solution: results from rock analogue experiments. J. Geophys.Res. Solid Earth 113 (B4), B04204.

Niemeijer, A., Marone, C., Elsworth, D., 2010. Frictional strength and strain weak-ening in simulated fault gouge: competition between geometrical weakeningand chemical strengthening. J. Geophys. Res. 115 (B10), B10207.

Odling, N.E., Harris, S.D., Knipe, R.J., 2004. Permeability scaling properties of faultdamage zones in siliclastic rocks. J. Struct. Geol. 26.

Olierook, H.K.H., Timms, N.E., Hamilton, P.J., 2014. Mechanisms for permeabilitymodification in the damage zone of a normal fault, northern Perth Basin,Western Australia. Mar. Pet. Geol. 50, 130e147.

Peach, C.J., Spiers, C.J., 1996. Influence of crystal plastic deformation on dilatancyand permeability development in synthetic salt rock. Tectonophysics 256 (1e4),101e128.

Rathbun, A.P., Marone, C., 2010. Effect of strain localization on frictional behavior ofsheared granular materials. J. Geophys. Res. 115, B01204. http://dx.doi.org/10.1029/2009JB006466.

Renard, F., Dysthe, D., Feder, J., 2001. Enhanced pressure solution creep ratesinduced by clay particles: experimental evidence in salt aggregates. Geophys.Res. Let. 28 (7), 1295e1298.

Revil, A., Grauls, D., Br�evart, O., 2002. Mechanical compaction of sand/clay mixtures.J. Geophys. Res. 107 (B11), 2293. http://dx.doi.org/10.1029/2001JB000318.

Rowan, M., Jackson, M., Trudgill, B., 1999. Salt-related fault families and fault weldsin the northern gulf of Mexico. AAPG Bull. 9 (9), 1454e1484.

Rutter, E.H., 1983. Pressure solution in nature, theory and experiment. J. Geol. Soc.140 (5), 725e740. http://dx.doi.org/10.1144/gsjgs.140.5.0725.

Rutter, E.H., Maddock, R.H., Hall, S.H., White, S.H., 1986. Comparative microstruc-tures of natural and experimentally produced clay-bearing fault gouges.PAGEOPH 124 (1e2), 3e30.

Samuelson, J., Elsworth, D., Marone, C., 2009. Shear-induced dilatancy of fluid-

B.M. Kaproth et al. / Marine and Petroleum Geology 78 (2016) 222e235 235

saturated faults: experiment and theory. J. Geophys. Res. 114 (B12), B12404.Savage, H.M., Brodsky, E.E., 2011. Collateral damage: evolution with displacement of

fracture distribution and secondary fault strands in fault damage zones.J. Geophys. Res. 116.

Scuderi, M.M., Kitajima, H., Carpenter, B.M., Saffer, D.M., Marone, C., 2015. Evolutionof permeability across the transition from brittle failure to cataclastic flow inporous siltstone. Geochem. Geophys. Geosys. 16 http://dx.doi.org/10.1002/2015GC005932.

Seebeck, H., Tenthorey, E., Consoli, c, Nicol, A., 2015. Polygonal faulting and sealintegrity in the Bonaparte Basin, Australia. Mar. Pet. Geol. 60, 120e135.

Shimamoto, T., 1986. Transition between frictional slip and ductile flow for haliteshear zones at room temperature. Science 231 (4739).

Shipton, Z.K., Evans, J.P., Thompson, L.B., 2005. The geometry and thickness ofdeformation-band fault core and its influence on sealing characteristics ofdeformation-band fault zones. In: Sorkhabi, R., Tsuji, Y. (Eds.), Faults, Fluid Flow,and Petroleum Traps. AAPG Memoir, vol. 85, pp. 181e195.

Skurtveit, 2013. Experimental Investigation of Deformation Mechanisms DuringShear-Enhanced Compaction in Poorly Lithified Sandstone and Sand.

Smith, D., 1966. Theoretical considerations of sealing and non-sealing faults. AAPG

Bull. 50 (2), 363e374.Sternlof, K.R., Chapin, J.R., Pollard, D.D., Durlofsky, L.J., 2004. Permeability effects of

deformation band arrays in sandstone. AAPG Bull. 88 (9), 1315e1329.Takahashi, M., Mizoguchi, K., Kitamura, K., Masuda, K., 2007. Effects of clay content

on the frictional strength and fluid transport property of faults. J. Geophys. Res.112 (B8), B08206.

Tueckmantel, T., Fisher, Q.T., Knipe, R.J., Lickorish, H., Khalil, S.M., 2010. Fault sealprediction of seismic-scale normal faults in porous sandstone: a case studyfrom the eastern Gulf of Suez rift, Egypt. Mar. Pet. Geol. 27, 334e350.

Wibberley, C.A.J., Shimamoto, T., 2003. Internal structure and permeability of majorstrike-slip fault zones: the median tectonic line in mie prefecture, southwestJapan. J. Struct. Geol. 25, 59e78.

Wong, T.F., Baud, P., 2012. The brittle-ductile transition in porous rock: a review.J. Struct. Geol. 44, 25e53.

Zhang, S., Tullis, T.E., 1998. The effect of fault slip on permeability and permeabilityanisotropy in quartz gouge. Tectono. 295 (1e2), 41e52.

Zhang, S., Tullis, T.E., Scruggs, V.J., 1999. Permeability anisotropy and pressure de-pendency of permeability in experimentally sheared gouge materials. J. Struct.Geol. 21 (7), 795e806.